• The Vendor's Bazaar forum is for commercial offers and transactions. Only unmoderated members can post here.

    diyAudio provides this forum for the convenience of our members, but makes no warranty nor assumes any responsibility. We do not vet any members. Use of this facility is at your own risk. Customers can post any issues in those threads as long as it is done in a civil manner. All diyAudio rules about conduct apply and will be enforced.

Output transformers for DACs

I have been using those transformers for a while in an ancient Sony CDP that contains a CXD2552Q2 D/A chip - I know, old technology - but I have always liked the big soundstage created by that player. Once I started tinkering with my gear I tried a variety of CDP mods - different capacitors, bypassing the output stage after the opamps - and, as a matter of course, different opamps - all the way to discrete ones (manufactured by two different companies).
Well, to cut a long story short - I was introduced to Ivan's transformers in a friend's project - and as it takes a bit longer than expected to replicate it - I decided to use the transformers in the above mentioned CDP - and - there is no way that I shall ever go back to any opamp - the only way that comes to mind to describe what the transformers have done to the music is - taking away any digital artifacts (for the lack of a better word). CDs - especially well-recorded ones - sound simply more natural without any digital "harshness" - and those that are less well recorded? Well, they sound the way they are, but still a lot better than before - and some even far less thin that I had always believed. So, yes, I am a very happy customer and the next step is going to be getting those trannies that can be used with a TDA1541 player.
 
Can one transformer be configured to provide balanced or single ended output or do you have to choose?

Sure, I can make 1:0.5+0.5 or 0.5+0.5:0.5+0.5 (if the DAC have differential output and center point is needed) version of turns ratio for your choice. You will connect coils different ways to provide SE or BAL outputs. It is extra 10USD to the cost of standard (1:1) version.

dubai2000 Simple and understandable 🙂 Thank you!
 
In case of ES90x8 chips there is no need to use grounded center point between + and - of the signal. Moreover my expiriences says me that avoiding tighting somehow to ground of the output signal from DAC (if it is possible) is better for SQ. There are some DACs that have + and - of the signal with respect to the ground (like PCM1792) - and in that case there will be no sound at all, untill you not connected to the ground. Say if you can avoid summing stage then no need to split the load. So in case of ES90x8 or AK449x - there is no need to "watch the ground". You can but not necessary.
 
Hi, I decided to go for 1:4 because balanced isnt needed that much and continuity of a single winding is preferable, also with half the step up for balanced it wouldnt sound the same.

Can you explain why 1:4 is preferred than 1:1 with lower I/V resistor?
 
...continuity of a single winding is preferable...
Yes, it is for the customers choice.

...also with half the step up for balanced it wouldnt sound the same.
The whole sound amplitude will be splitted for two waves. That is it. I believe there is no any impact to the sound quality inside my transformers while splittering coils because of toroidal type of the core I use in my transformers like it happenes in non-symmetrical types of cores.
Can you explain why 1:4 is preferred than 1:1 with lower I/V resistor?
In case of ES90x8 it is prefered to use primary coils with as lower active (DC) resistance as possible because I believe the Voltage mode of ES90x8 is only for sound checking (works/not), but the only really high SQ mode is the current mode.
In terms of low DCR the primary coil in my transformer I can make is about 20H. In this case 1:1 will be quiter than 1:4.
1:1 is better than 1:4 in terms of transformers specs (FR, slewrate etc.) but the louder sound level you can grab from your DAC without FR losses the better will be SQ (as far from noise flor as possible). So 1:4 is a good compromise between DCR/winding length and resulting FR and amplitude level.
 
They arrived today,
I have been trying to understand the whole thing better since

this didnt fully answer my question and im quite curious:
So 1:4 is a good compromise between DCR/winding length and resulting FR and amplitude level.
So the 1:4 transformer is trading current for voltage,
the 8k I/V resistor is converting current to voltage (i.e amplitude).
Wouldnt a 1:1 transformer with 2k resistor result in the same voltage level
with better specs of 1:1 transformer?

Is is that you are aiming for the ''sweet spot'' with transformer roll-off around 20Khz instead of the widest FR, avoiding filtering after the transformer?

Also should I naturally use higher (32k) i/v resistor with 9038pro (4x output) and 1:4?
as you say it will require some tweaking but that would be logical starting place?
 
Last edited:
i think I am confusing the impact of the I/V resistor value, in that higher or lower resistance will not result in more or less voltage linearly...

The question remains that if 4 times the current with Pro version will result in 4 times the voltage why not use 1:1?
 
Laserscrape, if the sensetive of the next stage (amplifier) is ~2V and the impedance is high (~50k for example) to achieve full power, then your impedance (resistor) value at the secondary should be 1.8kOhm in total. So, try to use 1k-2k shunt to be synchronised with your amp impedance.
 
Last edited:
Wouldnt a 1:1 transformer with 2k resistor result in the same voltage level
with better specs of 1:1 transformer?
The DC resistance of the 1:1 transformer used with AK4490 (for example) tens times more and sounds worse with ES.
Is is that you are aiming for the ''sweet spot'' with transformer roll-off around 20Khz instead of the widest FR, avoiding filtering after the transformer?
There is more powerfull LPF in case of ES (or AK) DACs - its inner Digital LPF. My aim is not to add additional LPF if possible.
Also should I naturally use higher (32k) i/v resistor with 9038pro (4x output) and 1:4? as you say it will require some tweaking but that would be logical starting place?
Yes, you can, but the output level will be respectively high.
[/QUOTE]
 
i think I am confusing the impact of the I/V resistor value, in that higher or lower resistance will not result in more or less voltage linearly...
Because there is the limitation by your next stage' impedance. If you will load transformer to the really high impedance like oscilloscope, then you will see the wave level dependance from resistor value adequatelly.

The question remains that if 4 times the current with Pro version will result in 4 times the voltage why not use 1:1?

But what if the next stage (amplifier) will be with 1kOhm of input impedance? In our days - it is not so rare case. I think it is not absolutely right to limit output capability of the powerful DAC when it can be more universal for different types of amplifiers including low input impedance models.
 
Oh, so i had the I/V resistors effect backwards! thats a relief.
With that in mind 2k was still high in volume, 500r is giving about the same output as the old op amp output stage, I use headphones so output level is probably very low.
thanks for helping me out.

Anyway the SQ is instantly appreciatable, very nice and natural sound 🙂


About 1:1 transformer, this is what confused me:
In terms of low DCR the primary coil in my transformer I can make is about 20H. In this case 1:1 will be quiter than 1:4.

The 1:1 you mention here wont have low DCR suitable for Sabre DAC?

EDIT: sorry, you just explained that
 
Last edited: